I'm not saying that the Linux Mark Institute is doing this
Good, because they aren't. The sublicense agreement only gives LMI the right to terminate if the sublicensee starts using the mark with unauthorized goods/services (something completely unrelated to Linux), or if they violate the agreement in another way.
The agreement doesn't give LMI the right to terminate just because the sublicensee smeared Linux's name. LMI cannot hold the company to any standards, codes of conduct, etc.
Usually these have been problems fixed by swapping hardware around until it works.
I'm not paid to spend hours debugging the kernel, I'm paid to get things working.
I wasn't even aware the kernel had a bugzilla until this thread. Last I heard Linus didn't want a bugzilla and wanted people to send shit to LKML.
Usually the way it goes down is the vendor kernel breaks, so we go to the vanilla kernel, that breaks, then we move hardware around until it works. You seem to be advocating using the vanilla kernel. That's not what many other replies to this are saying.
They are saying that specifically because when Linus decided to make the stable kernel unstable, he said the distros would make it stable. I don't agree with this philosophy one bit.
I can only think many of these problems are unforseen interactions between certain chipsets and add-in cards, because often changing the motherboard with a different one shows both the motherboard and the hardware work fine, just not together.
the company I work for is a "Microsoft Preferred Partner"
Why would you want to work at such a place?
The way I look at it, if a company is so hung up on credentialism, I don't want to work for them anyway.
Works out great. They are doing me a service, filtering themselves out of prospective places to work that don't share at least values in the same ballpark as mine.
You still haven't explained why the question whether you distribute pristine upstream sources or patched ones should even matter.
Quit harping on the LMI agreement and look at trademark law.
The LMI agreement lets you distribute modified versions, but trademark law does not does not give you that right in absence of such an agreement.
Lets assume The Gimp gets a trademark on that name.
If I take the Gimp and distribute it, everything is peachy. I can call it the Gimp. Without a license. Because it is the Gimp. Trademark lets you refer to the original product by its trademark, this is called nominative use.
But if I modify the Gimp, it's no longer "The Gimp", it's something else entirely. Something that if I were to distribute it as "The Gimp" could cause customer confusion and dilution of "The Gimp" as a trademark. I could be sued. I would lose. I have to buy a license from The Gimp Licensing Institute.
You see my point now? Modification and distribution of a product revokes your nominative "fair use" rights under trademark law.
This is an extremely important concept, I want to make sure people understand it.
Quoting again from the license agreement because so seem to think you can ignore it.
I can ignore it. Because I'm not talking about it. I'm talking about trademark law in general now, not LMI. I'm talking about what you are or aren't allowed to do without buying a license.
And the truth of the matter is, LMI could sue you for distributing "Linux Powered" things that were not pristine sources, if you didn't buy the license.
You have no nominative fair use defense, that went out the window when you modified it. You have no statutory fair use defense. You have no comparative advertising fair use defense. You have no parody fair use defense.
You have no defense. You can be sued. You will lose.
Given, the adaptec cards giving trouble were somewhat older, 40 and 80 meg cards. They still used aic-7xxx.
As for the 3com issue, it only showed up after several weeks of heavy traffic. The network would slow down to about half speed. Reloading the kernel module made it go back to full speed. I suspect some hard to track down memory leak.
"he" is gender neutral. It has been for years. Same with "man/men". Unless you think we should strip women of constitutional rights because the constitution doesn't mention women.
As I also mentioned in my other long post, the only reason I brought this up was because that was one of the arguments for LMI that I saw, that it would somehow prevent smear-journals with Linux in their names.
either that, or you're a troll, or maybe you have the WORST reading comprehension skills I ever encountered
Cool it on the ad hominems, I'm not trolling.
"Linux kernel with modifications" is explicitly allowed.
Which is why I said it does little to protect the Linux mark from damage. Which is the supposed reason for all this. BadCompany could still license the Linux mark, fuck up the kernel badly, and release thier new "linux product".
Nothing LMI could do would stop this after the fact. There's no termination clause for such actions.
The same is true of BadLinuxJournal. With something like a magazine, there's no software being distributed at all. There's no way to revoke the license if they are just a MS shill trying to smear Linux, the magazine is still about Linux, and is an "Authorized Service", it "documents Linux". No recourse once the sublicense is issued.
I saw this "stopping bad magazines" argument put forward as a reason for LMI too, yet it is completely impotent at stopping such things.
The BSD clause was deemed unsustainable because it became clear that F/OSS was going to be constructed from a great number of components of different copyright holders. This does not hold true in the same way for trademarks.
Yet. Once other major software projects realize they can scam money this way, we'll see lots more jumping on the bandwagon. Better make some room for a few hundred trademark acknowledgements, and keep a lot of extra money in the budget, once all the software projects demand money for the use of their name.
"But", I can hear you saying, "if you just use the project name to refer to the project itself, it's within trademark to use it that way without restriction". That's true. But what the distros distribute is rarely the original product. They are in fact distributing something different, something that isn't quite "The Gimp" anymore.
It wouldn't be a hard case for say The Gimp, to argue that what Red Hat is distributing isn't really "The Gimp" but in fact is a derived version that is apt to cause consumer confusion about what "The Gimp" really is. Thus Red Hat can't use the Gimp name under that provision, and must license the mark to use it with their product.
Yes, LMI allows derived works to be called "Linux" (if you pay their license).. but that very idea of derived works gives the trademark holder that much more power to control the use of their trademark, to dictate that it only be used when referring to the official releases.
If this sounds outlandish, it's exactly what happened with Red Hat and Centos. Red Hat has told Centos they may not refer to Red Hat Linux anywhere in their web site, documentation, etc.
Altavista was just as good as Google. For a while. Then Go bought them (or they bought Go or whatever). And Altavista started looking like the craptacular "portal" that for some reason these idiots thought users wanted.
The altavista started all these "online catalog of everything" ties. I don't know if anyone remembers the labyrinth of thousands and thousands of catalog pages on altavista, where you could buy anything, from shoes and sweaters to tents and food. This was around 2000.
Anyway, Altavista failed because they lost focus. As google tacks more and more functions on, they very well may fall victim to the same fate.
I'd say this entirely depends on the license agreement.
It's published on the site! I suggest you read it before commenting! It even includes an unsustainable "advertising clause", like the old BSD license that was thrown out.
Of course under threat of legal action, lika all contracts
No I mean suits for infringement which pretty much implies no contract.
Yes there is: Linus can simply not grant or revoke the license.
The sublicense agreement doesn't give them any permission to revoke the license just because the licensee starts doing something they don't agree with.
I know all about Della Croce. And for the last 9 years it was fine and good without these fees. Now suddenly they are putting the screws on, under threat of legal action.
Re:Interesting flash-based captcha
on
Defeating Captcha
·
· Score: 1
Plus, I thought the ADA applied mainly (only?) to government sites?
Probably... there is some court precedent that says some corporate web sites may not fall under the ADA. But it's not completely settled.
Just curious, how does 'chmod 000' prevent, for example, bots creating thousands of free email acounts?
That's your problem. There are other ways than captchas. Like one guy suggested, there are plenty of very simple questions that it is hard for a computer to solve, that don't rely on vision.
I'm all for more creative solutions, but not ones that lock out some small percentage of users for whatever reason.
Re:...the same features we delivered seven years a
on
Windows 95 Turns 10
·
· Score: 2, Interesting
Slashdot Mirror
Well, yes, but that's not a flaw in the encryption, that's a flaw in the idea of DRM.
Current encryption technology was never meant to be used this way.
What you are saying is like saying "All screwdrivers that are used as cold chisels will break eventually." Well, duh.
You can't know if that's true, unless you can prove P=NP.
How will buying more hardware now prevent laws from being passed in the future which outlaw the possession of devices without built-in DRM?
Laws passed in the future??
The DMCA bans it now, if it can be used to somehow circumvent existing DRM.
Yeah, that's almost as bad as an ISP using something like SPEWS. If I ever got an ISP that used SPEWS, they'd get a nice lawsuit.
You better tell all those people that don't understand trademark law that.
All the idiots that keep asserting you need no sublicense to say "Powered by Linux" are so full of shit.
At least we have a tangible example now, thanks.
I'm not saying that the Linux Mark Institute is doing this
Good, because they aren't. The sublicense agreement only gives LMI the right to terminate if the sublicensee starts using the mark with unauthorized goods/services (something completely unrelated to Linux), or if they violate the agreement in another way.
The agreement doesn't give LMI the right to terminate just because the sublicensee smeared Linux's name. LMI cannot hold the company to any standards, codes of conduct, etc.
Usually these have been problems fixed by swapping hardware around until it works.
I'm not paid to spend hours debugging the kernel, I'm paid to get things working.
I wasn't even aware the kernel had a bugzilla until this thread. Last I heard Linus didn't want a bugzilla and wanted people to send shit to LKML.
Usually the way it goes down is the vendor kernel breaks, so we go to the vanilla kernel, that breaks, then we move hardware around until it works. You seem to be advocating using the vanilla kernel. That's not what many other replies to this are saying.
They are saying that specifically because when Linus decided to make the stable kernel unstable, he said the distros would make it stable. I don't agree with this philosophy one bit.
I can only think many of these problems are unforseen interactions between certain chipsets and add-in cards, because often changing the motherboard with a different one shows both the motherboard and the hardware work fine, just not together.
Maybe the declaration of independence would have been a better example, if less legally relevant.
the company I work for is a "Microsoft Preferred Partner"
Why would you want to work at such a place?
The way I look at it, if a company is so hung up on credentialism, I don't want to work for them anyway.
Works out great. They are doing me a service, filtering themselves out of prospective places to work that don't share at least values in the same ballpark as mine.
Don't call it a day yet.
You still haven't explained why the question whether you distribute pristine upstream sources or patched ones should even matter.
Quit harping on the LMI agreement and look at trademark law.
The LMI agreement lets you distribute modified versions, but trademark law does not does not give you that right in absence of such an agreement.
Lets assume The Gimp gets a trademark on that name.
If I take the Gimp and distribute it, everything is peachy. I can call it the Gimp. Without a license. Because it is the Gimp. Trademark lets you refer to the original product by its trademark, this is called nominative use.
But if I modify the Gimp, it's no longer "The Gimp", it's something else entirely. Something that if I were to distribute it as "The Gimp" could cause customer confusion and dilution of "The Gimp" as a trademark. I could be sued. I would lose. I have to buy a license from The Gimp Licensing Institute.
You see my point now? Modification and distribution of a product revokes your nominative "fair use" rights under trademark law.
This is an extremely important concept, I want to make sure people understand it.
Quoting again from the license agreement because so seem to think you can ignore it.
I can ignore it. Because I'm not talking about it. I'm talking about trademark law in general now, not LMI. I'm talking about what you are or aren't allowed to do without buying a license.
And the truth of the matter is, LMI could sue you for distributing "Linux Powered" things that were not pristine sources, if you didn't buy the license.
You have no nominative fair use defense, that went out the window when you modified it.
You have no statutory fair use defense.
You have no comparative advertising fair use defense.
You have no parody fair use defense.
You have no defense. You can be sued. You will lose.
Given, the adaptec cards giving trouble were somewhat older, 40 and 80 meg cards. They still used aic-7xxx.
As for the 3com issue, it only showed up after several weeks of heavy traffic. The network would slow down to about half speed. Reloading the kernel module made it go back to full speed. I suspect some hard to track down memory leak.
It's not FUD. I'm just reporting what I've run into.
Report the bugs and help the kernel developers find them.
If no one else is using adaptec scsi cards or 3com ethernet cards, then we have a problem!
Usually though the bugs have already been reported to LKML and the thread just died with no one doing anything about it.
When I can't find a stable kernel for one of my servers, it's a serious problem.
It's been hard to get long uptimes with 2.6... the network drivers are leaky/crash, SCSI support sucks.
It's just not been very hot.
It only outlaws lasers explicitly designed to damage sight or cause permanent blindness.
Temporary blinding.. while it's questionable if such is very possible without risk of permanant injury, isn't forbidden.
Nice Tesla Coil there. Now stop wasting my tax money on this BS.
As far as I can tell this guy just has some lasers and Tesla coils and "artist's renderings" of terrorists being struck down by sparks.
There's a fundamental problem yet to be addressed. It's extremely difficult to incapacitate someone without risking their life.
His vision of "Zapping the hostages with the kidnappers and sorting it out later" is scary!
That's the real risk of less-lethal weapons, they lend themselves to overuse.
"he" is gender neutral. It has been for years. Same with "man/men". Unless you think we should strip women of constitutional rights because the constitution doesn't mention women.
As I also mentioned in my other long post, the only reason I brought this up was because that was one of the arguments for LMI that I saw, that it would somehow prevent smear-journals with Linux in their names.
either that, or you're a troll, or maybe you have the WORST reading comprehension skills I ever encountered
Cool it on the ad hominems, I'm not trolling.
"Linux kernel with modifications" is explicitly allowed.
Which is why I said it does little to protect the Linux mark from damage. Which is the supposed reason for all this. BadCompany could still license the Linux mark, fuck up the kernel badly, and release thier new "linux product".
Nothing LMI could do would stop this after the fact. There's no termination clause for such actions.
The same is true of BadLinuxJournal. With something like a magazine, there's no software being distributed at all. There's no way to revoke the license if they are just a MS shill trying to smear Linux, the magazine is still about Linux, and is an "Authorized Service", it "documents Linux". No recourse once the sublicense is issued.
I saw this "stopping bad magazines" argument put forward as a reason for LMI too, yet it is completely impotent at stopping such things.
The BSD clause was deemed unsustainable because it became clear that F/OSS was going to be constructed from a great number of components of different copyright holders. This does not hold true in the same way for trademarks.
Yet. Once other major software projects realize they can scam money this way, we'll see lots more jumping on the bandwagon. Better make some room for a few hundred trademark acknowledgements, and keep a lot of extra money in the budget, once all the software projects demand money for the use of their name.
"But", I can hear you saying, "if you just use the project name to refer to the project itself, it's within trademark to use it that way without restriction". That's true. But what the distros distribute is rarely the original product. They are in fact distributing something different, something that isn't quite "The Gimp" anymore.
It wouldn't be a hard case for say The Gimp, to argue that what Red Hat is distributing isn't really "The Gimp" but in fact is a derived version that is apt to cause consumer confusion about what "The Gimp" really is. Thus Red Hat can't use the Gimp name under that provision, and must license the mark to use it with their product.
Yes, LMI allows derived works to be called "Linux" (if you pay their license).. but that very idea of derived works gives the trademark holder that much more power to control the use of their trademark, to dictate that it only be used when referring to the official releases.
If this sounds outlandish, it's exactly what happened with Red Hat and Centos. Red Hat has told Centos they may not refer to Red Hat Linux anywhere in their web site, documentation, etc.
.5% of your gross revenue isn't a "nominal fee".. That's a big chunk.
Ahh, you forget the UI aspect.
Altavista was just as good as Google. For a while. Then Go bought them (or they bought Go or whatever). And Altavista started looking like the craptacular "portal" that for some reason these idiots thought users wanted.
The altavista started all these "online catalog of everything" ties. I don't know if anyone remembers the labyrinth of thousands and thousands of catalog pages on altavista, where you could buy anything, from shoes and sweaters to tents and food. This was around 2000.
Anyway, Altavista failed because they lost focus. As google tacks more and more functions on, they very well may fall victim to the same fate.
Do one thing, and do it well.
I'd say this entirely depends on the license agreement.
It's published on the site! I suggest you read it before commenting! It even includes an unsustainable "advertising clause", like the old BSD license that was thrown out.
Of course under threat of legal action, lika all contracts
No I mean suits for infringement which pretty much implies no contract.
Yes there is: Linus can simply not grant or revoke the license.
The sublicense agreement doesn't give them any permission to revoke the license just because the licensee starts doing something they don't agree with.
I know all about Della Croce. And for the last 9 years it was fine and good without these fees. Now suddenly they are putting the screws on, under threat of legal action.
Plus, I thought the ADA applied mainly (only?) to government sites?
Probably... there is some court precedent that says some corporate web sites may not fall under the ADA. But it's not completely settled.
Just curious, how does 'chmod 000' prevent, for example, bots creating thousands of free email acounts?
That's your problem. There are other ways than captchas. Like one guy suggested, there are plenty of very simple questions that it is hard for a computer to solve, that don't rely on vision.
I'm all for more creative solutions, but not ones that lock out some small percentage of users for whatever reason.
How can a process have a version number?